Meta-mistake: are fragile meta-analyses in ophthalmology worth the high cost?

Abstract

We read with interest the article by Nanji et al., which details that meta-analyses in ophthalmology are frequently fragile, and their statistical significance hinges on the event status of very few patients [1]. We would like to add to this finding with information regarding the relevance of meta-analyses within ophthalmology and how we could improve their overall reliability (Table 1).

Table 1 List of recommendations to enhance the reliability and utility of meta-analyses in ophthalmology.

Full size table

Meta-analyses were originally written to compile data across publications with similar research aims, to increase robustness and reliability of scientific research as a whole. However, with the current ease of reviewing literature, abundance of analytical software to simplify this process, new statistical techniques to generate different types of meta-analyses, and the relative accessibility given that meta-analyses do not require a lab setting, the number of meta-analyses published across all fields have starkly increased especially during the COVID-19 pandemic [2].